Automatic radio map creation in a fingerprinting‐based BLE/UWB localisation system

Automated Calibration of RSS Fingerprinting Based Systems Using a Mobile Robot and Machine Learning

“…When it comes to localisation of a moving person, the median trajectory error is similar to that reported in [ 22 ] (0.4 vs. 0.42 m). It is worth noting that, in the referenced work, where the signatures were gathered with a hybrid BLE/UWB system and the resulting radio maps were personalised, the maps were interpolated using the GPR.…”

“…In [ 21 ], a separate UWB-based positioning system is used. The research presented in [ 22 ] uses a hybrid BLE/UWB system.…”

“…The data can also be used to fit the parameters of the radio propagation models [ 22 ]. The paper [ 24 ] presents an adaptive signal model fingerprinting (ASMF) algorithm, in which the parameters of path loss, fading and shadowing models along with the access points (AP) locations are estimated and used to interpolate the radio map.…”

“…After the initial localisation tests, the regions, where the localisation accuracy was the lowest are additionally sampled and the procedure repeated, until the accuracy is at an acceptable level. In the case of the UWB assisted solutions [ 22 ] (trajectory error in 0.42–0.52 m range), [ 21 ] (mean error in 0.72–0.85 m range), the number of gathered signatures is large and gathered by the person who later takes part in the tests. Additionally the tests in [ 21 ] were performed in a single room, which limited the negative impact of the through the wall propagation.…”

“…From the above works, only a few were conducted in a living environment or a similar cluttered area [ 13 , 22 , 33 ], or using robots [ 10 , 23 , 24 , 25 ]. These methods will be the main yardstick for the proposed solution’s accuracy.…”